主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

Chinese Journal of Management Science ›› 2024, Vol. 32 ›› Issue (5): 103-112.doi: 10.16381/j.cnki.issn1003-207x.2021.0547

Previous Articles     Next Articles

Stochastic Simulation Integrated Method for Multi-source Uncertain Information and Its Application

Lu Wang1,Pingtao Yi2(),Weiwei Li2   

  1. 1.School of Management, Shenyang University of Technology, Shenyang 110027, China
    2.School of Business Administration, Northeastern University, Shenyang 110167, China
  • Received:2021-03-19 Revised:2021-05-25 Online:2024-05-25 Published:2024-06-06
  • Contact: Pingtao Yi E-mail:ptyi@mail.neu.edu.cn

Abstract:

Under the increasingly complex and uncertain evaluation environment, the expression of uncertain information has been further developed and many uncertain theories and methodologies have been introduced into comprehensive evaluation problems.A comprehensive evaluation problem involves such issues that the experts with different knowledge backgrounds usually need the freedom to provide the opinions by their individual preferences and the development of social platform and search software has made it possible to obtain diversified information, such as the text evaluation information and fragment information that people leave on the website inadvertently. Besides, the absolute ranking, which means that an alternative is superior to its next adjacent one with 100% probability, is lack of explanation for comprehensive evaluation problems containing multi-source information especially uncertain information. It is a meaningful and urgent issue to propose a novel method to integrate multi-source inputs to a reasonable output.For the sake of the problems above, a multi-source uncertain information (MSUI) integration framework was built to fuse all kinds of information mentioned in this paper by simulation techniques. Then, obtain the most likely ranking with pairwise priority probabilities. Specifically, the first problem is fusing the MSUI. The MSUI is normalized into a unified scope where the random numbers considering membership degree are generated by a certain distribution. The second problem is integrating the MSUI. The MSUI is classified by the types into different clusters, based on which the MSUI integration framework is established. The third problem is obtaining a reasonable output. the priority comparison of each alternative can be done by each simulation. After adequate simulation, the ratio of the times that one alternative is prior to another one to the total simulation times tends to be stable. Then, the pairwise priority matrix (PPM) is obtained. Based on the method of regression tree, the most likely ranking with pairwise priority probabilities can be obtained from the matrix.An application example that a company evaluates the comprehensive ability of 8 employees in the marketing department shows that: (1) The development of uncertain theory, such as fuzzy sets, linguistic information, allows the experts to describe multi-attribute evaluation problems more freely but more precisely. (2) The MSUI is fully tapped by adequate simulation, which avoids the employees being sorted by just one comparison. (3) After comparing the most likely ranking and the absolute ranking, the employee o4 is not absolutely superior to employee o5, but superior to employee o5 by 70.65% probability. The most likely ranking provides more information about the comparison among the employees.The main contributions of the method proposed in this paper are summarized as follows. (1) The fusion of MSUI ensures the information characteristics and fully mines the information value. (2) This information fusion platform has strong compatibility with fragment information, which expands the range of information available for evaluation problems. (3) Some possible outputs are obtained which differ from the absolute results by most methods, which can better explain the practical phenomenon in the actual world such as the weak team winning the strong one in a game.

Key words: comprehensive evaluation, multi-source uncertain information, information integrated framework, stochastic simulation, possibility ranking

CLC Number: